[Bioperl-l] Issue with Bio::SearchIO::psl (was: Bioperl bug 1977)

Chris Fields cjfields at uiuc.edu
Fri Mar 31 08:40:21 EST 2006


I'll try it with Mac OS X this weekend to confirm; I'm running v.  
10.4.5 with perl 5.8.6.

I noticed that there's no tests for psl in SearchIO.t which should  
have caught this error.  I'll double check that in case I'm  
mistaken.  If not, I'll add a few to see what happens... maybe we'll  
get some responses back?

I'll also forward this to the mail list to see if anybody else has  
had this issue.

Chris

On Mar 31, 2006, at 3:53 AM, Albert Vernon Smith wrote:

> Running your same code, on the same file, I get:
>
> Output:
> -------
> /usr/local/blat/db/hg17/hg17.2bit:chr5
>     100.00
> /usr/local/blat/db/hg17/hg17.2bit:chr21
>     90.00
> /usr/local/blat/db/hg17/hg17.2bit:chr5
>     85.00
> /usr/local/blat/db/hg17/hg17.2bit:chr13
>     80.00
> /usr/local/blat/db/hg17/hg17.2bit:chr7
>     80.00
> Use of uninitialized value in pattern match (m//) at /Users/albert/ 
> Documents/CSHL/cvswork/bioperl-live/Bio/SearchIO/psl.pm line 173,  
> <GEN1> line 10.
> --------
>
> This is current CVS, and I see the problem on Mac OS X, as well as  
> on Linux.
>
> As it stands the code for Bio::Search::psl *should* be fine (as I  
> run it in my head :-), and the error message is kinda weird.  The  
> last line is #10, so there should be a value for the line, unless  
> things are trying to cycle back over it again for some reason.
>
> -albert
>
>
> On 30.3.2006, at 22:10, Chris Fields wrote:
>
>> I'm running off bioperl-live from CVS (updated yesterday) and I get
>> everything to work on this end (no errors) using your file,  
>> although I'm
>> just printing names and HSP scores out, like this:
>>
>> --------------------------------------
>>
>> my $parser = Bio::SearchIO->new(-verbose    => $v,
>>                                 -file       => 'psl.out',
>>                                 -format     => 'psl');
>>
>> while (my $result = $parser->next_result) {
>>     while (my $hit = $result->next_hit) {
>>         print $hit->name,"\n";
>>         while (my $hsp = $hit->next_hsp) {
>>             print "    ",$hsp->score,"\n";
>>         }
>>     }
>> }
>>
>> --------------------------------------
>> Output:
>> --------------------------------------
>> /usr/local/blat/db/hg17/hg17.2bit:chr5
>>     100.00
>> /usr/local/blat/db/hg17/hg17.2bit:chr21
>>     90.00
>> /usr/local/blat/db/hg17/hg17.2bit:chr5
>>     85.00
>> /usr/local/blat/db/hg17/hg17.2bit:chr13
>>     80.00
>> /usr/local/blat/db/hg17/hg17.2bit:chr7
>>     80.00
>> --------------------------------------
>>
>> Is this a recent update of Bioperl?  There were several updates in  
>> CVS to
>> Bio::SearchIO::psl for various bugfixes over the last year,  
>> including one
>> that postdates the 1.5.1 release.  I would recommend trying the  
>> CVS version
>> (copy it over the your old version if possible or just install  
>> bioperl-live
>> from CVS).  If this doesn't work could you send your script?  It  
>> may be a
>> specific method that's acting up.
>>
>> Christopher Fields
>> Postdoctoral Researcher - Switzer Lab
>> Dept. of Biochemistry
>> University of Illinois Urbana-Champaign
>>
>>
>>> -----Original Message-----
>>> From: Albert Vernon Smith [mailto:smithav at cshl.edu]
>>> Sent: Thursday, March 30, 2006 1:49 PM
>>> To: Chris Fields
>>> Subject: Re: Bioperl bug 1977
>>>
>>> [Message never went out before.  Was stuck in outbox.]
>>>
>>> I've attached an output which causes issues.  While parsing this
>>> output gives me an issue, I'm actually doing something slightly
>>> different.  I have a webBlat server, and am getting output via
>>> LWP::UserAgent, and I take the psl returned from my query and pass
>>> that in memory (with IO::String) to the parser.  When I do that, I
>>> get a complaint which references the last line.  Still, parsing this
>>> as a file should be the same thing.
>>>
>>> -albert
>>
>>
>

Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign





More information about the Bioperl-l mailing list